Implications for future research

In this study, altered endocrine responses (i.e. delayed recovery) have been found both for the HPA axis as well as the sympathetic pathway (epinephrine). Whereas differences in self-reported distress and depressive symptoms between FR+ and FR- women did not reach

statistical significance, the endocrine alterations were detected despite the limitations in statistical power. There are at least two possible explanations for this difference. It is conceivable that the time window assessed by the measures of psychological distress was too small. The questionnaires used here are designed to measure stress symptomatology (GSI) and perceived stress (PSS) in the past week. Depressive symptoms (BDI) were measured during the past month. It can therefore not be ruled out that measures covering a larger interval of time might have shown a closer association with endocrine stress responses.

It is however also possible that the endocrine stress system is more sensitive to effects of familial breast cancer risk than self-report data. Apart from providing additional physiological information on FR+, this finding hints to a potential pathway leading to an increased familial risk.

In the following sections, possible implications of delayed endocrine stress recovery and heightened reactivity in women at high risk for breast cancer are discussed and possible future research strategies are outlined. It is appreciated that some of these implications are highly speculative at this point in an area where only little empirical evidence is available. However, future research may benefit from the integration of different fields of research as outlined below.

Stress, cortisol, cognition, and health behaviors

Screening visits for breast cancer have been found to be highly stressful experiences (Aro et al., 2000), especially in women who are scheduled for a further mammogram because of inconclusive, unsatisfactory, or abnormal results (Sandin et al., 2002). Such visits may be particularly stressful for women who are aware of their increased cancer risk. Beyond the potential effects of even further increased cancer risk based on immune and metabolic alterations associated with stronger endocrine stress responses, the observed phenomenon may also have effects relevant to behavioral variables. It is well known that acute release of glucocorticoids (e.g. cortisol) inhibits cognitive processes (see Lupien & McEwen, 1997).

Prolonged administration of glucocorticoids has also been found to inhibit cognitive function in healthy volunteers (Young et al., 1999). When given acutely, glucocorticoids significantly interfere with declarative and working memory in humans (e.g. Wolf et al., 2001).

Due to the increased cortisol response during such screening visits, FR+ women may potentially remember less information from the diagnostic interview and the screening guidelines provided by their physician. This could in turn affect compliance with surveillance suggestions and health behavior. According to the bio-behavioral model proposed by

Andersen et al. (1994), compliance and health behaviors are important factors influencing disease incidence as well as progression.

Endocrine responses in FR+ subgroups with strongly increased psychological distress.

As already mentioned above, with respect to the theoretical model proposed by Rees et al.

(2001), women’s personal experiences with a relative suffering from the disease may play a role in terms of their illness representations. It may be associated with increased psychological distress. It should thus further be studied if certain subgroups of FR+ women with a similar “objective” risk but different experiences differ on measures of perceived stress and endocrine stress responses. Zakowski et al. (1997) provided first evidence for a role of experience in psychological distress. Considering the human and animal endocrine literature, it appears likely that FR+ women with the experience of bereavement may exhibit even stronger endocrine acute stress response alterations. This warrants further investigation.

Additionally to the genetic breast cancer risk as established by family history assessment, these women may further be put at risk because of the alterations in acute endocrine stress responses. In this respect, it is of potential interest that similar epinephrine associations where found in women with higher scores on the BDI, even when statistically controlling for FR+.

This suggest that for a subgroup of FR+ women who exhibit more depressive symptoms, detrimental effects of overly strong stress responses may even be stronger. To date, mostly cross-sectional data is available in support of McEwen’s theory of allostatic load (McEwen &

Seeman, 1999). Evidence from prospective studies to investigate the predictive value of endocrine stress response alterations as reported here is therefore clearly needed.

Genetic vs. environmental cause?

From the results of the study presented here, it is not possible to draw a conclusion about the underlying mechanism leading to increased endocrine responses and delayed recovery. As noted above, the familial disposition and the altered endocrine responses could be a result of the same or a related genetic disposition. Furthermore, in this study, we could not show a direct association of psychological distress and endocrine responses. The phenomenon may represent a genetic or a “acquired” hyper-activity of the HPA axis and the sympathetic pathway. It could also be the effect of a synergistic interaction of genetic preposition and stress. In any case, there is experimental evidence that HPA hyperactivity is progressive over time. That is, frequent activation of the HPA axis and chronically elevated glucocorticoid levels may (further) enhance hyperactivity of this system. Chronic stress has been found to

lead to glucocorticoid receptor downregulation, especially type I (MR) receptors in the hippocampus, resulting in a hyposuppressive state in rats (e.g. Young et al. 1990; Mizoguchi et al., 2001). It has further been demonstrated that chronic stress paradigms lead to downregulation of MR and GR mRNA expression (Francisca et al., 1996; Herman et al., 1995). Stress-induced glucocorticoid elevations have been linked to neuronal loss in the hippocampus due to the neurotoxic effects (Stein-Behrends et al., 1994). The hippocampus, where type I receptors (MR) are densely expressed, is considered to play a crucial role in the negative feedback loop of the HPA axis (DeKloet et al., 1998). It is tempting to speculate that heightened and longer-lasting glucocorticoid stress responses further increase HPA hyperactivity via hippocampal neurodegeneration. The causing factor as well as the development over time of endocrine response patterns remains to be elucidated in future studies.

Are endocrine stress response alterations involved in the psycho-biological interaction causing increased cancer incidence in FR+ women?

Above, the findings of this study have been discussed with respect to what the causing factor of the endocrine alterations may be (genetic vs. psychological/environmental). It has also been mentioned that, according to McEwen’s theory, these alterations (through the effects and outcomes of the primary mediators) may lead to tertiary outcomes, including cancer (McEwen

& Seeman, 1999, p. 40). It appears possible that the failure to shut-off acute stress responses is not merely an epiphenomenon of FR+ associated psychological distress (“acquired hyperactivity”) but that it may actually represent one possible pathway linking familial breast cancer risk and breast cancer incidence.

What is known is that a family history of breast cancer is associated with increased likelihood for developing the disease (Pharoah et al., 1997). It also seems clear that while a number of gene mutations likely to determine genetic susceptibility have been identified, up to half of the hereditary breast cancers occur in the absence of such mutations. It has also been reported that mutation carriers are not necessarily affected by the disease (see section 2.1.1.). This suggests that carrying the gene mutation does not inevitably lead to disease and other factors may also play a role in determining whether or not a women will be affected by breast cancer.

It is therefore tempting to speculate that some of the hereditary breast cancers are actually caused by a HPA hyperresponsivess, which may in part be genetically determined. As noted above, there is some evidence that such a genetic determination of endocrine stress responsiveness exists.

The study presented here provides the first preliminary evidence that FR+ women show altered endocrine stress responses closely resembling type III allostatic load as described by McEwen. Preliminary experimental evidence suggests that the HPA axis and the sympathetic pathway play a role in modulating physiological functions of suspected importance in tumor control. These include poorer repair of DNA and increases in the frequency of sister chromatid exchange, and NK cell function. Below, I will briefly highlight some examples. For instance, Shakkar and Ben-Eliyahu (1998) reported that in vivo administration of an adrenergic agonist (metaproterenol, MP) resulted in decreased NK cytotoxicity in a rat model.

Furthermore, MP administration lead to a 10-fold increase in tumor cells metastases when a NK-sensitive tumor model was used. Adrenergic antagonists showed the opposite effects.

DNA repair mechanisms and sister chromatid exchange have also been reported to be affected by stress in animal models (e.g. Glaser et al., 1985; Fischman & Kelly, 1987). The mechanism linking the physiological changes induced by stress and the observed changes in DNA repair and sister chromatid exchange is not known. Kiecolt-Glaser and Glaser (1999) suggest that

“one or more stress hormones may mediate these responses” (p.1605), since the HPA axis and the autonomic nervous system are activated by stress.

It can be hypothesized that endocrine stress response alterations are one possible working pathway linking family history and breast cancer incidence. Future research will have to test whether such a hypothesis can be supported empirically. It would primarily be necessary to show the predictive value of endocrine alterations for cancer incidence independent of BRCA1 and BRCA2 mutations.